Background: Suicidal behavior in late-life is heterogeneous and likely include multiple pathways. We propose that age at first suicide attempt carries useful information for the clinical, cognitive, and broader biological heterogeneity of suicidal behavior and could help identify subgroups of distinct risk profiles. The presentation will show a series of studies to discuss the current state of research behind the suggested early- late-onset categorization of late-life suicidal behavior. Methods: Studies used subsamples of a larger sample collected by the Pittsburgh Longitudinal Research Program in Late-life Suicide. Participants had unipolar, non-psychotic major depression, it included attempters N=245, mean age 64.1 years (SD=9.53), 53% female, and a demographically similar depressed non-attempter comparison group N=286, mean age 64.9 (SD=8.61), 52% female. Statistical approach: While our earlier studies used sample median to establish the cutoff age for early- vs. late-onset suicidal behavior, our most recent studies used Gaussian mixture modelling of age of onset of suicidal behavior and found a bimodal distribution, verified by bootstrapping, with a statistically optimal cut-off around 30 years of age at first attempt. Results: Early-onset suicide attempters exhibited significantly higher levels of neuroticism, introversion, and cluster B traits, were more likely to suffer childhood abuse, and more frequently had co-morbid PTSD than those who first attempt suicide in late life. Only early-onset, but not late-onset, suicide attempters had higher levels of familial and social transmission of suicidal behavior compared to non-suicidal depressed elderly. We found that late-onset attempters are largely indistinguishable from non-suicidal depressed participants with the exception of cognitive functioning. Late-onset suicidal behavior is associated with a broad range of cognitive deficits, including memory, processing speed, and executive dysfunction, possibly signaling a dementia prodrome. In contrast, both early-onset and late-onset suicide attempters display impairment in cognitive control (including conflict monitoring and cognitive flexibility). The personality trait conferring increased vulnerability to late-onset suicidal behavior was high orderliness, a subcomponent of conscientiousness. Conclusion: Our studies showed multidimensional heterogeneity of risk factors for suicidal behavior based on age of onset. Older adults with early-onset attempts display life-long vulnerabilities that are similar to young suicide attempters (e.g., family history of suicidal behavior, emotion dysregulation), while late-onset attempters display age-specific vulnerabilities such as worse cognitive function that could hinder adaptation to age-related stressors.